Advanced composite structures, such as foam core carbon fiber reinforced polymer composites, are increasingly being
used in applications which require high strength, high in-plane and flexural stiffness, and low weight. However, the
presence of in situ damage due to manufacturing defects and/or service conditions can complicate the failure
mechanisms and compromise their strength and reliability. In this paper, the capability of detecting damages such as
delaminations and foam-core separations in X-COR composite structures using non-destructive evaluation (NDE) and
structural health monitoring (SHM) techniques is investigated. Two NDE techniques, flash thermography and low
frequency ultrasonics, were used to detect and quantify the damage size and locations. Macro fiber composites (MFCs)
were used as actuators and sensors to study the interaction of Lamb waves with delaminations and foam-core
separations. The results indicate that both flash thermography and low frequency ultrasonics were capable of detecting
damage in X-COR sandwich structures, although low frequency ultrasonic methods were capable of detecting through
thickness damages more accurately than flash thermography. It was also observed that the presence of foam-core
separations significantly changes the wave behavior when compared to delamination, which complicates the use of wave
based SHM techniques. Further, a wave propagation model was developed to model the wave interaction with damages
at different locations on the X-COR sandwich plate.
We describe a stochastic ltering approach for tracking progressive fatigue damage in structures, wherein physically based damage evolution information is combined with active sensing guided wave measurements. The input waveform used to excite dispersive modes within the structure is adaptively con gured at each time step in order to maximize the damage estimation performance. The damage evolution model is based on Paris Law, and hidden Markov modeling of time-frequency features obtained from received signals is used to de ne the measurement model. Damage state estimation is performed using a particle lter. Results are presented for fatigue crack estimation in an aluminum specimen.